Here is how I see it, Microsoft would want to hold onto their dominance in the enterprise, from where something like 80% of their revenue comes. To follow the market, they need to develop a viable Cloud Computing system((It doesn’t really matter whether it is the future or not, they HAVE to put something into this space)). The biggest challenge is to develop a system that Corporate customers will buy.
Lets assume that this is true. Corporate customers are very unlikely to buy Azure anytime soon. It needs to be proven with large workloads, applications need to be specifically developed for it, and data centers need to be built that host the exact servers, storage and networks that this will require.
Normally, a radical product like this would mean that a sales force would be setup to target early adopters and vertical applications and deparately try to get them to buy it. In return, customers will get a LOT of help, access to internal tech support, free marketing and, possibly, bragging rights as a market leader. The Azure concept is too big to find an early adopter that would be credible.
The other criteria is that the “market” expects to Microsoft to make a “Cloud Computing” play. Here it is, but with options to sell direct to Enterprise in the future, a two way bet.
So how do you handle these customer objections ? Today, you go the Web 2.0 way: put it on the Internet, give it away for free for the first couple of years while you iron out the bugs and scaling problems. You refine the feature sets according to customer feedback, and develop your operational tools for miantenance of the system. You might make a bit of cash back in advertising revenue or by charging a small fee for premium services, but the real winner is when you are ready to sell it to Enterprises.
During this time, developers have been using the Internet Cloud to develop applications and test them, some corporates might even have started using these new applications on the cloud. Enough, at least, to prove that they work and provide marketing collateral for white papers.
At this point, Enterprise IT staff will be ready to consider Azure for Misty Computing.
Misty Computing is where you can see through the cloud to the elements that make it up such as servers, software, networking: a deterministic platform. They are all joined together by a software system that automates provisioning and operation but the IT Manager will be able to see what the system contains and make deterministic decisions about its future.
Cloud Computing is intangible and unimaginably impossible for more Corporates, Misty Computing offers local, quasi-visible, controlled resources that have the advantages of the cloud.
Cloudy Computing will need a lot of iron in the early days, but I feel confident that this will shrink rapidly – so much so that Misty Computing will be the preferred alternative.
Is Cloud Computing just an beta version of the next generation data centers ?